Knowing the relationships among individuals in structured societies appears to require a mental representation of the social network, one that categorizes group members on multiple dimensions, such as family relationship, history of interactions, or dominance status. In large groups, this complex category structure must be assembled on the basis of partial information, using cognitive capabilities analogous to transitive inference, and it must be flexible enough to be rapidly revised and updated as circumstances change. Such capabilities are integral components of human intelligence and are strongly affected by abnormal development or neural disease. Given their importance, however, this aspect of social cognition is surprisingly poorly understood. The research we propose uses an established, robust model system, a highly social bird that has shown great expertise in inferring their own dominance status from the interactions of others and in assembling and re-assembling hierarchies of arbitrary stimuli based on partial information. Emphasis will be placed on understanding the cognitive structures used in representing stimulus hierarchies, their operation in inferring relationships among members of a social group, and the role of the hippocampal formation in enabling such relational representations. The proposed studies will extend previous research efforts along three convergent lines. The first will use operant techniques to examine the cognitive mechanisms supporting transitive inference in social and non-social species. The second will use observations of social groups and of pairs of birds in staged encounters to explore natural, social inference during dominance interactions. The third will search for the neural substrates involved in relational encoding of social hierarchies, examining the effects of lesions of the hippocampal formation. By integrating these three approaches, the proposed research provides a unique opportunity to test hypotheses about the mechanism of social cognition as well as to evaluate the relationship between natural social cognition and performance on analogous laboratory tests. The methodology will be readily applicable to the study of transitive inference and social cognition in humans, providing significant benefits for the diagnosis and treatment of disorders of cognitive development, such as autism, amnesia, and schizophrenia. PUBLIC HEALTH RELEVANCE: A variety of mental disorders, including autism, schizophrenia, and Alzheimer's disease, degrade the patients' ability to understand and make inferences about social relationships. By integrating operant, observational, and neurobiological approaches, the proposed research provides a unique opportunity to investigate the mechanism of social cognition directly and to evaluate the relationship between natural social cognition and performance on laboratory tests. The methodology will be applicable to the study of transitive inference and social cognition in humans, providing significant benefits for the diagnosis and treatment of cognitive disorders.